How eCommerce Uses Big Data – High-Value Use Cases to Consider

Did you know that only around 1% of generated digital data is being analysed, even though 53% of the companies (interviewed in the Big Data Analytics Market Study) use big data analytics?

The analytical capabilities big data brings to eCommerce are instrumental. In 2019, 100% of the eCommerce enterprises in Thailand utilised big data analytics for creating a new product. With big data analytics, companies can create new products that provide increased consumer value and at the same time, connect with their desires.

Using predictive analytics, companies can forecast the post-launch performance of a product and determine the best marketing strategies and optimal production and distribution chains for rapid growth.

In other words, big data in Ecommerce is revolutionising how companies identify new means to increase their sales and sustain in an increasingly competitive landscape. From aggregating and analysing customer data such as browsing history, shopping preferences, and social media activity businesses are becoming more aware of their customers’ needs.

What is Big Data in eCommerce?

Big data refers to the massive amounts of data generated every second from various sources such as social media, online transactions, research, and more. This data can be structured, semi-structured or unstructured are is characterised by what is known as the four Vs: volume, velocity, variety and veracity.

In the eCommerce industry, big data is used to help brands and retailers make informed decisions, design tailored products and services, foster innovation and observe market trends. Big data eCommerce engages customers in two primary ways: providing personalised recommendations and increasing the visibility of a product’s/service’s information.

Sources of big data in eCommerce

Here are some common sources of big data in eCommerce:

  • Transactional data: This source contains vital information such as the buyer’s ID and product’s ID, transaction time and data, personal information, and a sum of money your customer spent on purchase through a given payment method. As such, transactional data provides deep insight into a customer’s behaviour and product/service demand.
  • Social data: This source includes metadata from social media (e.g. Facebook, Twitter/X, Instagram etc.) such as location, hashtags, likes, shares, and records of brand interactions. Social data is used to provide more personalised online shopping experiences for customers as well as for target audience segmentation.
  • Third-party data: This source comes from the integration of third-party programs. Third-party data is useful for providing information on market trends, goods and services representation and pricing policies, giving businesses new opportunities to expand.

Benefits of Using Big Data in eCommerce

1. Cost reductions

With big data analytics, companies can save costs by optimizing their supply chain and reverse logistics (product returns) process, reducing costs in marketing, and even analysing the purchase-return data to determine products that are more likely to be returned.

2. Better Customer Experience and Satisfaction

With 80% of customers preferring personalised shopping experiences, big data can help businesses enhance satisfaction rates by taking into account multiple factors such as purchase history, preferences, browsing habits, demographics, and polarity.

3. Better marketing strategies

Big data analytics makes it easier to see which marketing strategies performed better and their impact on overall sales. Likewise, with techniques such as sentiment analysis, marketers can better understand the emotional feedback of customers and tweak their campaigns accordingly.

4. Ability to predict trends

By accumulating data from multiple sources such as social media networks, customer reviews and feedback, and weather conditions, big data analytics can help companies predict trends, even before they hit the mainstream.

5. Price optimisation

Since pricing is a critical factor for driving purchase decisions, companies need to always have a competitive advantage by offering the best price. Big data analytics will optimise this process by considering multiple parameters such as competitors’ prices, profitability, weather conditions and the uniqueness of the product to determine the best price.

6. Security

With fraudulent claims and payment frauds ruling the digital world, eCommerce businesses are vulnerable to losing money. Big data analytics can help organisations monitor atypical spending behaviour and predict even before an individual can cause damage.

How Does eCommerce Use Big Data?

Let’s explore some big data eCommerce case studies and see how the biggest companies use big data in eCommerce to improve their practices.

1. Personalised experience

9 out of 10 marketers imply that customers crave for personalization, and 4 in 10 use machine learning for that. Personalisation alone increases conversion rates by 63%. The most prominent role of big data in eCommerce is the power of personalisation it offers to organisations.

Almost every online store uses personalization to drive sales, especially eCommerce giants like Amazon, Walmart, and Target. With the extensive use of predictive analytics for personalization and recommendation engines, Target is so advanced that some customers get creeped out by the suggestions they receive.

Disturbingly, there’s an incident in which Target figured out a teen girl was pregnant, even before her father got to know. To avoid that level of creepiness, Target started to mix irrelevant ads with personalised ads to make the whole process look a little random.

Personalised coupons from Target Corporation. Image Credit:

To help with its personalised customer experience, Target assigns a Guest ID number to every customer. This ID contains every bit of information used for personalisation, including, name, email address, credit card details, demographics, and purchase history.

2. Customer analysis

With a diverse group of customers to cater to, businesses will have to define multiple target groups before running each marketing campaign. Thanks to big data, customer segmentation is easier than ever as the more you know about a customer, the more precisely you can group them.

Segmenting customers based on purchase data. Image Credit:

Similarly, not all customers are created equal. A minority of your customers will be responsible for the majority of your profits – just like the Pareto Principle. Segmentation is also crucial to identify customers who bring in the most value (customer lifetime value modelling) – thereby allowing you to channel your efforts accordingly.

Sentiment analysis dashboard. Image Credit:

Another great instance of using big data in eCommerce is sentiment analysis. Sentiment analysis helps businesses understand the emotional feedback of customers towards products and services – thereby allowing companies to tweak their deliverables. Customer data from sources such as social media networks (eg., Twitter, Facebook, Instagram etc.) is used for this.

Take the case of BikeBerry, for example. Being an online bicycle and accessories retailer, BikeBerry was always concerned about customer retention and engagement.

For that reason, they “thoughtlessly” engaged in discount-driven retention campaigns via emails, following a one-discount-fits-all approach that was ultimately harming their budget.

BikeBerry hoped to maximise the utilisation of their customer retention budget by sending offers to only those customers who needed a “push” (in the form of coupons) to make a purchase.

But this also meant that BikeBerry was also sending coupons to those customers who would have purchased without the coupons. To better optimise their retention campaigns, BikeBerry used big data analytics and tools to differentiate between the customers who are more likely to purchase without incentives and the ones who rely on them.

And the results were almost instantaneous. Their email marketing campaigns saw a sudden increase of 133% in sales, along with a 200% increase in user activity. The rate of returning customers doubled, and they started to spend at least 30% more than before.

Their big data tools work by aggregating and analysing multiple data sets, including purchase history, behavioural data, browsing patterns, demographics, email opening rates and time. BikeBerry now runs a tight ship by sending timed emails only when a customer is most likely to open them.

3. Customer experience

Customer experience and service can make or break an eCommerce business. 90% of Americans see customer service as a deciding factor to remain loyal to a business. If customers aren’t satisfied with your service, 13% of them will tell other people how unhappy they are. Likewise, 72% of customers, if they had a positive experience, will share it with other people.

With big data in eCommerce, companies can deliver multi-channel customer support with decreased response time and increased problem-solving efficiency.

North Face, an American activewear product company is a prime example of using big data in eCommerce for enhanced customer experience.

The company teamed up with IBM to utilise the potential of Watson, a program that combines artificial intelligence (AI), machine learning (ML)and big data analytics to deliver the right experience to customers.

Conversing with Watson on North Face’s website. Image Credit:

Customers can directly engage with Watson, like its a human salesperson and ask questions or input preferences to receive customised recommendations. The answers a customer provides in the initial phase will be stored and used to shape future conversations with Watson.

This also means that a customer’s conversation will shape the suggestions Watson makes and the recommendation engine of the website. Watson also takes into account the purchase history, browsing activities, and customer reviews from social media networks to enhance digital shopping experiences.

4. Secure online payments

With the introduction of mobile-commerce (use of handheld devices like smartphones and tablets to perform online purchases and transactions), enterprises have to provide support for multiple payment methods – which invariably increases the number of associated threats. Analysing big data can help businesses identify anomalies and detect fraud.

Predictive analytics can also allow eCommerce companies to identify potential threats and initiate preventive strategies for the same. A notable example of this is how PayPal (which processed $US712 billion in 2019 alone) uses big data analytics for fraud detection.

An illustration of rule-based fraud detection. Image Credit:

Paypal’s big data engine collects more than 20 terabytes of log data every day. With machine learning algorithms, the system can compare patterns and identify fraudulent transactions.

Paypal also uses logistic regression (a statistical model that estimates the probability of an event to occur based on previous data), and advanced techniques like gradient-boosted trees (a type of machine learning boosting, which relies on the intuition that the previous models, when combined with the next best model, minimises overall prediction errors) to make its machine learning algorithms more accurate.

5. Price optimisation

When it comes to customer purchase decisions, pricing is one of the most prominent factors. Along with better customer service and product quality, lower pricings offered by competitors can be one another reason why customers may feel tempted to churn and break loyalty.

Example of Price vs. Demand, Sales revenues, and Gross profits, for one product. Image credit:

Price optimisations can be overwhelming for humans to perform as the associated data is enormous. Several factors, such as competitor analysis, weather conditions, product availability, and cost-to-profit ratio, must be taken into account. Fortunately, big data analytics makes the entire process a piece of cake.

An excellent example of this is Agoda, an online booking agency. One of the primary factors that contributed to Agoda’s popularity is its competitive price guarantee. This means, that if you book a hotel room through Agoda and get a cheaper offer on some other (genuine) website, the agency will either match the rate or beat it.

But, returning the money, once paid, can be tiresome. There’s a lot of paperwork, processing fees, and extra hassle associated with returns. To prevent all that, Agoda extensively uses big data to predict the lowest possible price at the right time.

For that, Agoda takes into account multiple factors such as competing prices (both online and offline), weather trends, calendar events, news-related events, labour issues, features of the hotel and the room etc.

6. Demand forecasting

Demand forecasting with the help of predictive analytics is a notable form of big data application in the eCommerce industry. By implementing statistical models and machine learning algorithms on historical data, big data systems can predict demands and help businesses fine-tune their supply chain processes.

For example, take the case of Amazon. When Jeff Bezos registered the orders, delivered the packages to the post office and tracked the inventory – all by himself – demand forecasting may have been relatively simpler. Fast forward 25 years and Amazon has more than 200 million unique visitors per month.

Along with demand forecasting, Amazon also uses big data analytics for price optimization, its anticipatory shipping model and its recommendation engine.

A brief illustration of how Amazon’s demand forecasting works. Image Credit:

In 2019, Amazon created the Galaxy data lake, a centralised repository that stores and manages massive amounts of unstructured and structured data with speed and accuracy. Such a data lake is crucial for Amazon as delayed data can lead to millions of dollars in lost revenue.

These data lakes are powerful foundations for AI and ML, which further shape Amazon’s demand forecasting models. Amazon also takes into account multiple factors, including social media trends, economic viewpoints, website traffic logs, historical sales, seasonal demands, technological advancements, and price of goods to better forecast demands.

7. Supply chain management

It’s safe to say that without big data analytics, supply chain management would have been an arduous responsibility. Big data makes inventory management efficient, even with the ever-increasing number of products. Big data also plays a pivotal role in fleet management and eliminating instances such as overstocking.

Walmart is a prime example of companies using big data in eCommerce processing 2.5 petabytes of data (unstructured) every hour.

Unsurprisingly, Walmart’s reach means that it receives millions of customers per day. Such a huge volume means that having an efficient supply chain management system is critical for a smooth operation.

For that, Walmart makes extensive use of big data analytics for supply chain management and delivering personalised shopping experiences and recommendations.

Benefits of using big data analytics in supply chain management. Image Credit:

To ensure smooth supply chain management, Walmart Labs (a subsidiary of Walmart) uses big data analytics across multiple supply chain activities, including sourcing, shipment preparation, transportation, last-mile scheduling/routing and pick-up.

For example, to show the estimated delivery date, known as delivery promise, Walmart considers multiple factors such as:

  • Distance between the fulfilment centre and the customer.
  • Inventory levels of the item.
  • Available shipping methods and costs associated with each.
  • The capacity of shipping mode.

An example of route optimization algorithms at work (

Similarly, big data analytics allow Walmart to significantly reduce costs and time associated with shipment preparation.

By considering the distance between each item (ordered by a customer) in the warehouse, Walmart’s picking optimisation strategy will suggest the shortest or most efficient route for the picker to follow. Analytics is also helpful for packing optimisation – that is, choosing the correct box size for shipment.

Through the implementation of big data analytics into their supply chain, Walmart witnessed numerous sustainable competitive advantages such as reduced inventory carrying costs, lower product costs, competitive pricing, and more importantly, increased customer satisfaction.

Case study to understand data science application in eCommerce

In the health and beauty eCommerce sector, an innovative company leveraged AI for predictive replenishment and personalized cross-selling, enhancing customer engagement. The company integrated a predictive consumption model and an AI-powered recommendation engine to accurately forecast consumption rates and tailor product suggestions.

This approach reduced stockouts, improved customer satisfaction by 5-10%, and boosted sales by 10-30%. Additionally, it enhanced inventory management, leading to a 10-20% reduction in costs and improved cash flow, demonstrating the significant impact of AI and big data in retail.

To learn more about the effects of big data in eCommerce, read our case study on big data for eCommerce here.

Final Thoughts

By 2020, there will be at least two billion online shoppers – meaning, the data produced will be humongous in volume. Be it online or offline, businesses can use big data analytics to predict the needs and preferences of their customers, optimise the supply chain process for cost reduction and most importantly, improve sales by staying ahead of trends and demands.